Bull, Christopher Neil and Asfiandy, Dommy and Gledson, Ann and Mellor, Joseph and Couth, Samuel and Stringer, Gemma and Rayson, Paul Edward and Sutcliffe, Alistair Gordon Simpson and Keane, John and Zeng, Xiao-Jun and Burns, Alistair and Leroi, Iracema and Ballard, Clive and Sawyer, Peter Harvey (2016) Combining data mining and text mining for detection of early stage dementia : the SAMS framework. In: Resources and ProcessIng of linguistic and extra-linguistic Data from people with various forms of cognitive/psychiatric impairments (RaPID '16) workshop : Proceedings of the 10th International Conference on Language Resources and Evaluation (LREC '16). European Language Resources Association (ELRA), pp. 35-40.
combining_data_mining.pdf - Accepted Version
Available under License Creative Commons Attribution-NonCommercial.
Download (355kB)
Abstract
In this paper, we describe the open-source SAMS framework whose novelty lies in bringing together both data collection (keystrokes, mouse movements, application pathways) and text collection (email, documents, diaries) and analysis methodologies. The aim of SAMS is to provide a non-invasive method for large scale collection, secure storage, retrieval and analysis of an individual’s computer usage for the detection of cognitive decline, and to infer whether this decline is consistent with the early stages of dementia. The framework will allow evaluation and study by medical professionals in which data and textual features can be linked to deficits in cognitive domains that are characteristic of dementia. Having described requirements gathering and ethical concerns in previous papers, here we focus on the implementation of the data and text collection components.